-
监控中心接收到机器人回传的原始图像后,并不能直接显示巡检结果,不利于大批量的数据判断,因此我们为自主巡检系统提供了一套完善的识别方案,针对不同形式的设备采用不同的处理方式,提取其中的有用信息[17-18],完成对图像信息的识别。信息提取过程如图10所示。
对于指针式仪表,我们首先对原始图像去噪增强,并进行二值化处理,突出原始图像中的有用信息。若拍摄角度产生倾斜,我们还将进一步对图像进行倾斜矫正,消除拍摄角度带来的影响,如图11所示。
一般情况下,指针式表盘的刻度均匀分布,因此我们采用角度拟合的方式识别指针读数,如图12所示。
记表盘最小值、最大值所在位置分别为P1、P2,表针旋转轴所在位置为P3,表针方向所在的线段为L1,L2和L3分别为P1与P3、P2与P3的连线。L2和L3的夹角$ \alpha $以及L1和L2的夹角$ \beta $可以轻松获得。这样,表盘的读数问题就转变为表针指向的识别问题,只要表针指向识别准确,表盘读数V可通过式(1)计算得到:
$$ \begin{array}{c}V={V}_{{\rm{min}}}+\dfrac{\beta }{\alpha }\left({V}_{{\rm{max}}}-{V}_{{\rm{min}}}\right)\end{array} $$ (1) 式中:
${V}_{{\rm{min}}}$、${V}_{{\rm{max}}}$——仪表量程的最小值和最大值。
若表盘刻度之间呈现倍数的增长关系时,此方法仍然具有适用性,表盘读数V可通过式(2)计算得到:
$$ V = {M^{{V_{{\rm{min}}}} + \tfrac{\beta }{\alpha }\left( {{V_{{\rm{max}}}} - {V_{{\rm{min}}}}} \right)}} $$ (2) 式中:
M——刻度之间的倍数。
对于数字式仪表,我们训练了一个卷积循环神经网络对其读数进行识别。卷积循环神经网络,是目前较为流行的图文识别模型,此方法对不定长的文本序列进行识别时,不必先对单个文字进行切割,而是将文本识别转化为时序依赖的序列学习问题,就是基于图像的序列识别。
图13以英文单词“STATE”为例展示了卷积循环神经网络对文字的识别过程[19]。图像输入至神经网络后,首先使用卷积网络提取文本特征,随后利用深度双向LSTM网络对特征向量进行融合,提取字符序列的上下文特征,得到每列特征的概率分布,最后通过转录层进行预测得到最终的文本序列。
此外,我们还对阀门和带灯按钮的开关状态进行了识别。监控中心获得实际仪表读数和阀门、按钮的开关状态后,将其与设备阈值进行比对。当设备状态发生变化,如漏水、漏油、温度异常、设备附件变化时,发生异常报警并自动定位异常发生地点。操作人员可对异常信息进行审核确认,快速明确设备存在的问题,为电厂安全稳定运行提供保障。
巡检任务完成后将会生成当次的巡检报告,并储存在监控中心,供操作人员浏览确认。监控中心可对历次的巡检数据进行分析,衡量设备在一段时间内的运行状态,从而判断相应时间内设备运行是否正常。为专业人员全面掌握锅炉房、汽机房、升压站、辅助车间的运行状态提供可靠的依据。
Design and Development of 5G+ Robot Autonomous Patrol Inspection System in Intelligent Power Plant
-
摘要:
目的 为了改变目前以人工巡检为主的电厂巡检模式,提高电厂巡检作业的智能化水平,同时充分掌握机器人自主巡检系统的设计方法。 方法 文章从机器人自主巡检系统的设计思路出发,明确系统功能,分别从硬件架构和软件功能两方面给出巡检系统的设计方案。并以此设计方案为依据,设计出一台集先进性与实用性于一体的5G+机器人自主巡检系统。此外,为使得机器人的巡检结果能够更加直观地反映出电厂的实际运行情况,文章基于智能图像处理技术,完成了对仪表读数、阀门、按钮开关状态等信息的分析处理。 结果 5G+机器人自主巡检系统能够实现地图构建、任务部署,并能自主导航执行巡检任务,自动识别巡检设备,同时利用5G回传巡检数据,完成巡检结果分析并自动生成巡检报告。 结论 开发的5G+智慧电厂机器人自主巡检系统可大幅提升巡检效率,提高电厂运维的智能化水平,为电厂的自主巡检作业提供了有力的技术支撑,为专业人员全面掌握锅炉房、汽机房、升压站、辅助车间的运行状态提供可靠的依据,在实现减员増效的同时保障了重点区域设备安全运行,更好地助力智慧电厂建设。 Abstract:Introduction In order to change the current patrol inspection mode of power plant mainly based on manual patrol ins-pection, improve the intelligent level of power plant patrol inspection, and fully grasp the design method of robot autonomous patrol inspection system. Method This paper started from the design idea of the robot autonomous patrol inspection system, defined the system functions, and gave the design scheme of the patrol inspection system from the hardware architecture and software functions. Based on this design scheme, a 5G+ robot autonomous patrol inspection system integrating progressiveness and practicality was designed. In addition, in order to make the robot's patrol inspection results more intuitively reflect the actual operation of the power plant, this paper, based on intelligent image processing technology, completed the analysis and processing of instrument readings, valves, button switch status and other information. Result The 5G+ robot autonomous patrol inspection system can realize map construction, task deployment, autonomous navigation and execution of patrol inspection tasks, automatic identification of patrol inspection equipment, and use 5G to return patrol inspection data, complete the analysis of patrol inspection results and automatically generate patrol inspection reports. Conclusion The 5G+ robot autonomous patrol inspection system in intelligent power plant developed in this paper can significantly improve the patrol inspection efficiency, improve the intelligent level of power plant operation and maintenance, provide strong technical support for the independent patrol inspection operation of the power plant, provide reliable basis for professionals to comprehensively grasp the operation status of the boiler room, steam turbine room, booster station and auxiliary workshop, and ensure the safe operation of equipment in key areas while reducing personnel and increasing efficiency, better assist the construction of smart power plants. -
[1] 黄山, 吴振升, 任志刚, 等. 电力智能巡检机器人研究综述 [J]. 电测与仪表, 2020, 57(2): 26-38. DOI: 10.19753/j.issn1001-1390.2020.002.005. HUANG S, WU Z S, REN Z G, et al. Review of electric power intelligent inspection robot [J]. Electrical measurement & instrumentation, 2020, 57(2): 26-38. DOI: 10.19753/j.issn1001-1390.2020.002.005. [2] 郑师应, 李源, 杨博涵, 等. 5G+行业现场网技术与产业发展综述 [J]. 电信科学, 2022, 38(增刊1): 17-27. DOI: 10.11959/j.issn.1000-0801.2022104. ZHENG S Y, LI Y, YANG B H, et al. Technology evolution and industry development of 5G+ industrial proximity network [J]. Telecommunications science, 2022, 38(Suppl. 1): 17-27. DOI: 10.11959/j.issn.1000-0801.2022104. [3] 刘智勇, 赵晓丹, 祁宏昌, 等. 新时代无人机电力巡检技术展望 [J]. 南方能源建设, 2019, 6(4): 1-5. DOI: 10.16516/j.gedi.issn2095-8676.2019.04.001. LIU Z Y, ZHAO X D, QI H C, et al. Prospect of UAV power inspection technology in new era [J]. Southern energy construction, 2019, 6(4): 1-5. DOI: 10.16516/j.gedi.issn2095-8676.2019.04.001. [4] 孙仝, 薛菲. 基于机载激光LiDAR的电网巡检应用研究 [J]. 南方能源建设, 2018, 5(3): 133-139. DOI: 10.16516/j.gedi.issn2095-8676.2018.03.022. SUN T, XUE F. Research and application of power grid inspection based on airborne LiDAR system [J]. Southern energy construction, 2018, 5(3): 133-139. DOI: 10.16516/j.gedi.issn2095-8676.2018.03.022. [5] 张雪焱, 伦振坚, 侯婷, 等. ±800 kV柔直换流站全封闭阀厅智能巡检系统研究 [J]. 南方能源建设, 2019, 6(4): 144-151. DOI: 10.16516/j.gedi.issn2095-8676.2019.04.023. ZHANG X Y, LUN Z J, HOU T, et al. Research on intelligent inspection system in enclosed valve hall of ±800 kV VSC-HVDC converter station [J]. Southern energy construction, 2019, 6(4): 144-151. DOI: 10.16516/j.gedi.issn2095-8676.2019.04.023. [6] 尚文静. 智慧电厂建设管理平台的设计与思考 [J]. 中国新通信, 2022, 24(13): 28-30. DOI: 10.3969/j.issn.1673-4866.2022.13.011. SHANG W J. Design and consideration of the construction management platform for smart power plants [J]. China new telecommunications, 2022, 24(13): 28-30. DOI: 10.3969/j.issn.1673-4866.2022.13.011. [7] 张燕东, 田磊, 李茂清, 等. 智能巡检机器人系统在火力发电行业的应用研发及示范 [J]. 中国电力, 2017, 50(10): 1-7. DOI: 10.11930/j.issn.1004-9649.201706125. ZHANG Y D, TIAN L, LI M Q, et al. Application and development of intelligent inspection robot system in thermal power plant [J]. Electric power, 2017, 50(10): 1-7. DOI: 10.11930/j.issn.1004-9649.201706125. [8] 黎睿. 面向电力巡检机器人的SLAM算法研究与系统设计 [D]. 杭州: 浙江大学, 2017. LI R. Research and system design of SLAM algorithm for power inspection robot [D]. Hangzhou: Zhejiang University, 2017. [9] 邓爱林, 冯钢, 刘梦婕. 5G+工业互联网的关键技术与发展趋势 [J]. 重庆邮电大学学报(自然科学版), 2022, 34(6): 967-975. DOI: 10.3979/j.issn.1673-825X.202106020189. DENG A L, FENG G, LIU M J. Key technologies and development trends in 5G+ industrial internet [J]. Journal of Chongqing University of posts and telecommunications (natural science edition), 2022, 34(6): 967-975. DOI: 10.3979/j.issn.1673-825X.202106020189. [10] 吴建清, 宋修广. 同步定位与建图技术发展综述 [J]. 山东大学学报(工学版), 2021, 51(5): 16-31. DOI: 10.6040/j.issn.1672-3961.0.2021.168. WU J Q, SONG X G. Review on development of simultaneous localization and mapping technology [J]. Journal of Shandong University (engineering science), 2021, 51(5): 16-31. DOI: 10.6040/j.issn.1672-3961.0.2021.168. [11] 吴功平, 杨智勇, 王伟, 等. 巡检机器人自主充电对接控制方法 [J]. 哈尔滨工业大学学报, 2016, 48(7): 123-129. DOI: 10.11918/j.issn.0367-6234.2016.07.020. WU G P, YANG Z Y, WANG W, et al. On auto-docking charging control method for the inspection robot [J]. Journal of Harbin institute of technology, 2016, 48(7): 123-129. DOI: 10.11918/j.issn.0367-6234.2016.07.020. [12] 毛军, 付浩, 褚超群, 等. 惯性/视觉/激光雷达SLAM技术综述 [J]. 导航定位与授时, 2022, 9(4): 17-30. DOI: 10.19306/j.cnki.2095-8110.2022.04.003. MAO J, FU H, CHU C Q, et al. A review of simultaneous localization and mapping based on inertial-visual-lidar fusion [J]. Navigation positioning and timing, 2022, 9(4): 17-30. DOI: 10.19306/j.cnki.2095-8110.2022.04.003. [13] 田野, 陈宏巍, 王法胜, 等. 室内移动机器人的SLAM算法综述 [J]. 计算机科学, 2021, 48(9): 223-234. DOI: 10.11896/jsjkx.200700152. TIAN Y, CHEN H W, WANG F S, et al. Overview of SLAM algorithms for mobile robots [J]. Computer science, 2021, 48(9): 223-234. DOI: 10.11896/jsjkx.200700152. [14] 李晓旭, 马兴录, 王先鹏. 移动机器人路径规划算法综述 [J]. 计算机测量与控制, 2022, 30(7): 9-19. DOI: 10.16526/j.cnki.11-4762/tp.2022.07.002. LI X X, MA X L, WANG X P. A survey of path planning algorithms for mobile robots [J]. Computer measurement & control, 2022, 30(7): 9-19. DOI: 10.16526/j.cnki.11-4762/tp.2022.07.002. [15] 詹京吴, 黄宜庆. 融合安全A*算法与动态窗口法的机器人路径规划 [J]. 计算机工程, 2022, 48(9): 105-112,120. DOI: 10.19678/j.issn.1000-3428.0062264. ZHAN J W, HUANG Y Q. Path planning of robot combing safety A* algorithm and dynamic window approach [J]. Computer engineering, 2022, 48(9): 105-112,120. DOI: 10.19678/j.issn.1000-3428.0062264. [16] 姚进鑫, 刘丽桑, 何栋炜, 等. 融合优化A*算法与动态窗口法的动态路径规划算法研究 [J]. 重庆理工大学学报(自然科学), 2022, 36(7): 197-207. DOI: 10.3969/j.issn.1674-8425(z).2022.07.025. YAO J X, LIU L S, HE D W, et al. Research on dynamic path planning algorithm based on optimized A* algorithm and dynamic window approach [J]. Journal of Chongqing University of technology (natural science), 2022, 36(7): 197-207. DOI: 10.3969/j.issn.1674-8425(z).2022.07.025. [17] 宋国栋. 煤矿电力巡检机器人设备在线识别方法研究与实践 [J]. 煤炭科技, 2022, 43(1): 100-104. DOI: 10.19896/j.cnki.mtkj.2022.01.020. SONG G D. Research and practice on online identification method of coal mine electric inspection robot equipment [J]. Coal science & technology magazine, 2022, 43(1): 100-104. DOI: 10.19896/j.cnki.mtkj.2022.01.020. [18] 张文杰. 基于图像配准与视觉显著性检测的指针仪表识别研究 [D]. 重庆: 重庆大学, 2016. ZHANG W J. Pointer meter recognition via image registration and visual saliency detection [D]. Chongqing: Chongqing University, 2016. [19] SHI B G, BAI X, YAO C. An end-to-end trainable neural network for image-based sequence recognition and its application to scene text recognition [J]. IEEE transactions on pattern analysis and machine intelligence, 2017, 39(11): 2298-2304. DOI: 10.1109/TPAMI.2016.2646371. [20] 权美香. 基于多传感器信息融合的单目视觉SLAM算法研究 [D]. 哈尔滨: 哈尔滨工业大学, 2021. DOI: 10.27061/d.cnki.ghgdu.2021.000085. QUAN M X. Research on multi-sensor information fusion based monocular visual slam algorithm [D]. Harbin: Harbin Institute of Technology, 2021. DOI: 10.27061/d.cnki.ghgdu.2021.000085.